<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Sofia Berg on AI Agents News</title>
    <link>https://aiagentsnews.top/authors/sofia-berg/</link>
    <description>Recent content in Sofia Berg on AI Agents News</description>
    <generator>Hugo</generator>
    <language>en</language>
    <lastBuildDate>Wed, 29 Jul 2026 01:16:56 +0000</lastBuildDate>
    <atom:link href="https://aiagentsnews.top/authors/sofia-berg/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Agent tools: Why 53 functions beat raw model size</title>
      <link>https://aiagentsnews.top/posts/agent-tools-why-53-functions-beat-raw-model-size/</link>
      <pubDate>Wed, 29 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/agent-tools-why-53-functions-beat-raw-model-size/</guid>
      <description>The APIBank benchmark tests 53 distinct tools to prove agent competence relies on tool access, not just parameter count or model size.</description>
    </item>
    <item>
      <title>Sovereign coding: How model flexibility protects data</title>
      <link>https://aiagentsnews.top/posts/sovereign-coding-how-model-flexibility-protects-data/</link>
      <pubDate>Sun, 26 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/sovereign-coding-how-model-flexibility-protects-data/</guid>
      <description>Over 2M installs show developers prefer sovereign coding where code stays local until explicit cloud selection via personal API keys.</description>
    </item>
    <item>
      <title>Inventory solutions: Blue Yonder vs o9 for retail planning</title>
      <link>https://aiagentsnews.top/posts/inventory-solutions-blue-yonder-vs-o9-for-retail-planning/</link>
      <pubDate>Sat, 25 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/inventory-solutions-blue-yonder-vs-o9-for-retail-planning/</guid>
      <description>Retailers using AI see 5, revenue growth while cutting operational costs by up to a significant portion, according to AllAboutAI data.</description>
    </item>
    <item>
      <title>Agent architecture: five components before scaling</title>
      <link>https://aiagentsnews.top/posts/agent-architecture-five-components-before-scaling/</link>
      <pubDate>Mon, 20 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/agent-architecture-five-components-before-scaling/</guid>
      <description>Avoid infinite loops by defining the five components an effective agent requires. Learn how retrieval, tools, and memory prevent chaotic outputs.</description>
    </item>
    <item>
      <title>Agent architecture beats raw model size today</title>
      <link>https://aiagentsnews.top/posts/agent-architecture-beats-raw-model-size-today/</link>
      <pubDate>Sun, 19 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/agent-architecture-beats-raw-model-size-today/</guid>
      <description>With 5 million weekly Codex users, agent architecture now dictates output quality more than raw model size or reasoning engines alone.</description>
    </item>
    <item>
      <title>Cline agent: open runtime for PlanAndAct workflows</title>
      <link>https://aiagentsnews.top/posts/cline-agent-open-runtime-for-planandact-workflows/</link>
      <pubDate>Sun, 19 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/cline-agent-open-runtime-for-planandact-workflows/</guid>
      <description>This open runtime unifies IDE and terminal workflows, boasting 64.2k GitHub stars while executing bash commands without vendor lock-in.</description>
    </item>
    <item>
      <title>Inference latency: Why 41B active beats 975B</title>
      <link>https://aiagentsnews.top/posts/inference-latency-why-41b-active-beats-975b/</link>
      <pubDate>Thu, 16 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/inference-latency-why-41b-active-beats-975b/</guid>
      <description>Learn why 41B active parameters matter more than total count for agent latency. Discover selection criteria for reliable, cost-effective systems.</description>
    </item>
    <item>
      <title>Agent memory types: Why most systems miss 5 of 7</title>
      <link>https://aiagentsnews.top/posts/agent-memory-types-why-most-systems-miss-5-of-7/</link>
      <pubDate>Tue, 14 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/agent-memory-types-why-most-systems-miss-5-of-7/</guid>
      <description>Most agents use only 2 of 7 memory types. Learn why working memory evaporates and how episodic storage fixes long-term agent autonomy.</description>
    </item>
    <item>
      <title>Agent workflows: Build a self-correcting course system</title>
      <link>https://aiagentsnews.top/posts/agent-workflows-build-a-self-correcting-course-system/</link>
      <pubDate>Tue, 14 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/agent-workflows-build-a-self-correcting-course-system/</guid>
      <description>Learn how 57% of firms now use multistep workflows to replace fragile prompts with a self-correcting Researcher and Judge pipeline.</description>
    </item>
    <item>
      <title>Agentic AI orchestration: avoid brittle systems</title>
      <link>https://aiagentsnews.top/posts/agentic-ai-orchestration-avoid-brittle-systems/</link>
      <pubDate>Tue, 14 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/agentic-ai-orchestration-avoid-brittle-systems/</guid>
      <description>The agentic AI market reaches $9.14 billion by 2026. Learn how state persistence and orchestration prevent brittle multiagent system failures.</description>
    </item>
    <item>
      <title>AI agent framework: LangGraph leads 18 deployments</title>
      <link>https://aiagentsnews.top/posts/ai-agent-framework-langgraph-leads-18-deployments/</link>
      <pubDate>Tue, 14 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/ai-agent-framework-langgraph-leads-18-deployments/</guid>
      <description>Analysis of 18 deployments shows LangGraph leads production readiness. Learn how to prevent context loss and infinite loops in your agent systems.</description>
    </item>
    <item>
      <title>Autonomous agent risks require sandboxed code</title>
      <link>https://aiagentsnews.top/posts/autonomous-agent-risks-require-sandboxed-code/</link>
      <pubDate>Tue, 14 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/autonomous-agent-risks-require-sandboxed-code/</guid>
      <description>By 2028, Gartner predicts a significant share of daily work decisions will be made autonomously by AI agents.</description>
    </item>
    <item>
      <title>Code workflow orchestration with Zencoder&#39;s multi-agent</title>
      <link>https://aiagentsnews.top/posts/code-workflow-orchestration-with-zencoders-multi-agent/</link>
      <pubDate>Tue, 14 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/code-workflow-orchestration-with-zencoders-multi-agent/</guid>
      <description>Zencoder collapses the cost of intelligence into a single subscription, granting access to every frontier model without per-token anxiety.</description>
    </item>
    <item>
      <title>crewAI 1.14.7: Secure State and Pluggable Backends</title>
      <link>https://aiagentsnews.top/posts/crewai-1147-secure-state-and-pluggable-backends/</link>
      <pubDate>Tue, 14 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/crewai-1147-secure-state-and-pluggable-backends/</guid>
      <description>crewAI 1.14.7 resolves critical CVEs and adds pluggable backends. Learn how state isolation supports the framework&#39;s 54.7k GitHub stars.</description>
    </item>
    <item>
      <title>crewAI update: Snowflake Cortex and trained agents</title>
      <link>https://aiagentsnews.top/posts/crewai-update-snowflake-cortex-and-trained-agents/</link>
      <pubDate>Tue, 14 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/crewai-update-snowflake-cortex-and-trained-agents/</guid>
      <description>crewAI 1.14.7a1 adds native Snowflake Cortex integration and lazy loading to cut startup times. Learn how trained agents persist state across 54.7k stars.</description>
    </item>
    <item>
      <title>Governed AI platforms fix context loss</title>
      <link>https://aiagentsnews.top/posts/governed-ai-platforms-fix-context-loss/</link>
      <pubDate>Tue, 14 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/governed-ai-platforms-fix-context-loss/</guid>
      <description>Brunelly replaces generic outputs with architecture-aware planning to fix fragmented engineering pipelines and restore lost system context.</description>
    </item>
    <item>
      <title>LangGraph 1.2.4: Why 14ms Latency Beats Runtime Crashes</title>
      <link>https://aiagentsnews.top/posts/langgraph-124-why-14ms-latency-beats-runtime-crashes/</link>
      <pubDate>Tue, 14 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/langgraph-124-why-14ms-latency-beats-runtime-crashes/</guid>
      <description>LangGraph 1.2.4 adds 14ms latency to prevent critical runtime failures. Learn to verify GPG signatures for commit 054a6f3 before deploying agents.</description>
    </item>
    <item>
      <title>LlamaAgents 0.0.14: Why Agents Are Now Microservices</title>
      <link>https://aiagentsnews.top/posts/llamaagents-0014-why-agents-are-now-microservices/</link>
      <pubDate>Tue, 14 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/llamaagents-0014-why-agents-are-now-microservices/</guid>
      <description>Version 0.0.14 of the llamaagents framework treats every agent as an independent microservice, removing the illusion of local execution for scalable systems.</description>
    </item>
    <item>
      <title>OpenAI Jalapeño chip cuts inference costs by half</title>
      <link>https://aiagentsnews.top/posts/openai-jalape%C3%B1o-chip-cuts-inference-costs-by-half/</link>
      <pubDate>Tue, 14 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/openai-jalape%C3%B1o-chip-cuts-inference-costs-by-half/</guid>
      <description>OpenAI&#39;s Jalapeño chip targets a 50% cost drop per token. We analyze the 3nm specs and what proprietary silicon means for your agent stack.</description>
    </item>
    <item>
      <title>Tool call requests: cut redundancy from 98% to 2%</title>
      <link>https://aiagentsnews.top/posts/tool-call-requests-cut-redundancy-from-98-to-2/</link>
      <pubDate>Tue, 14 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/tool-call-requests-cut-redundancy-from-98-to-2/</guid>
      <description>Optimized frameworks slash redundant tool calls from 98% to 2%. Learn how structured schemas and event-driven resumption build reliable AI agents.</description>
    </item>
    <item>
      <title>Tools for nondeterministic agents: fix 120s latency</title>
      <link>https://aiagentsnews.top/posts/tools-for-nondeterministic-agents-fix-120s-latency/</link>
      <pubDate>Tue, 14 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/tools-for-nondeterministic-agents-fix-120s-latency/</guid>
      <description>Agent requests often take 30 to 120 seconds, far exceeding the 50ms standard. Learn to build tool contracts that handle this nondeterministic latency safely.</description>
    </item>
    <item>
      <title>Agentic Patterns: Designing Reliable AI Systems</title>
      <link>https://aiagentsnews.top/posts/agentic-patterns-designing-reliable-ai-systems/</link>
      <pubDate>Sun, 12 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/agentic-patterns-designing-reliable-ai-systems/</guid>
      <description>Explore 1,672 commits of executable lessons on agent memory, context engineering, and securing systems against injection attacks.</description>
    </item>
    <item>
      <title>Production-ready Agents: Enforcing Deterministic State</title>
      <link>https://aiagentsnews.top/posts/production-ready-agents-enforcing-deterministic-state/</link>
      <pubDate>Sat, 11 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/production-ready-agents-enforcing-deterministic-state/</guid>
      <description>Seventy-seven percent of AI agent projects fail to reach production, leaving only a fraction of deployments operational according to recent 2026 data.</description>
    </item>
    <item>
      <title>Agent OS stops style drift in AI code</title>
      <link>https://aiagentsnews.top/posts/agent-os-stops-style-drift-in-ai-code/</link>
      <pubDate>Fri, 10 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/agent-os-stops-style-drift-in-ai-code/</guid>
      <description>Agent OS indexes existing repos to stop style drift, turning 170,000 GitHub stars into a system that enforces codebase standards before generation.</description>
    </item>
    <item>
      <title>Agent building: 150 lines of TypeScript code</title>
      <link>https://aiagentsnews.top/posts/agent-building-150-lines-of-typescript-code/</link>
      <pubDate>Thu, 09 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/agent-building-150-lines-of-typescript-code/</guid>
      <description>Build a functional research agent in 150 lines using strict TypeScript interfaces to prevent runtime errors and manage conversation history.</description>
    </item>
    <item>
      <title>AI coding agent explained: beyond autocomplete</title>
      <link>https://aiagentsnews.top/posts/ai-coding-agent-explained-beyond-autocomplete/</link>
      <pubDate>Thu, 09 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/ai-coding-agent-explained-beyond-autocomplete/</guid>
      <description>An AI coding agent plans multistep tasks, executes code, and iterates without handholding, moving far beyond simple autocomplete to true agency.</description>
    </item>
    <item>
      <title>OpenHands SDK: Build Agents That Run Terminal Commands</title>
      <link>https://aiagentsnews.top/posts/openhands-sdk-build-agents-that-run-terminal-commands/</link>
      <pubDate>Tue, 07 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/openhands-sdk-build-agents-that-run-terminal-commands/</guid>
      <description>The OpenHands SDK powers 2,013 commits of agent logic, enabling terminal execution and browsing within a single Python-defined agentic loop.</description>
    </item>
    <item>
      <title>Agent frameworks compared: LangGraph 0.2&#43; vs CrewAI 0.86&#43;</title>
      <link>https://aiagentsnews.top/posts/agent-frameworks-compared-langgraph-02-vs-crewai-086/</link>
      <pubDate>Mon, 06 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/agent-frameworks-compared-langgraph-02-vs-crewai-086/</guid>
      <description>Gartner predicts 40% of enterprise apps will use task-specific agents by 2027. Compare how LangGraph and CrewAI handle this complex shift.</description>
    </item>
    <item>
      <title>OpenHands Agent: Why EventSourced Beats Stateful</title>
      <link>https://aiagentsnews.top/posts/openhands-agent-why-eventsourced-beats-stateful/</link>
      <pubDate>Mon, 06 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/openhands-agent-why-eventsourced-beats-stateful/</guid>
      <description>OpenHands version 1.7.0 splits logic into modular packages, replacing the monolithic V0 design for better local deployment and audit trails.</description>
    </item>
    <item>
      <title>Agent personas boost value 23x over default settings</title>
      <link>https://aiagentsnews.top/posts/agent-personas-boost-value-23x-over-default-settings/</link>
      <pubDate>Sun, 05 Jul 2026 00:00:00 +0000</pubDate>
      <author>AI Agents News Editorial</author>
      <guid>https://aiagentsnews.top/posts/agent-personas-boost-value-23x-over-default-settings/</guid>
      <description>Configured AI agents drive a 23x value increase by stripping verbose output. Learn how specialized personas reduce token costs in complex development.</description>
    </item>
  </channel>
</rss>
